Odp: A few PIE questions

From: Piotr Gasiorowski
Message: 255
Date: 1999-11-13

 
----- Original Message -----
From: Gwydionash@...
To: cybalist@egroups.com
Sent: Saturday, November 13, 1999 7:27 AM
Subject: [cybalist] A few PIE questions

I would appreciate any information anyone could give me about the following 
questions relating to PIE grammar and phonetics.

1.  How did laryngeals affect roots that began with consonants?  For example, 
the PIE root *Hme becomes "eme", "emou", and "emoi" in the oblique cases of 
Attic Greek when used after prepositions (except for "pros me"), and the 
enclitics "me", "mou", and "moi" are used without prepositions.  Is this a 
PIE rule?  And are there rules governing when a larygeal affects an initial 
consonant and when it doesn't?

 
In PIE, word-initial larygeals before a consonant seem to have been non-syllabic, which means that no "schwa" developed in that position. The eventual loss of the laryngeals took place independently in various IE languages, hence their different behaviour in various branches. In Greek and Hittite (and also in Armainian) an inserted vowel normally breaks up laryngeal+nasal and laryngeal+liquid clusters, but the degree of stress certainly played a role, hence the clipped enclitic forms in Greek. Sanskrit (like most other languages) simply drops the laryngeal without a trace. The same differential treatment is also visible in rare triconsonantal onsets involving a laryngeal, as in the "star" word *xster-, which yields Greek aste:r, Arm. astÅ‚, vs. Latin stella, English star, Sanskrit st@... (Instr. pl.).
 

2.  How do IE linguists think the schwa was pronounced?  And did it tend to 
change to a common vowel in the daughter languages?  (I have noticed it 
changing to "a" a lot).

 
Most of the reflexes are indeed [a] or zero, but high vowels (Germanic *u, Aryan *i) also appear. Assuming that the schwa was phonemically different from *a (which some people have denied), the most likely value is a mid-central vowel (the unstressed vowel sound in many modern languages, including English). But the phonemicisation of schwa took place after the complete loss of the laryngeals, which occurred parallelly in different branches. It is quite possible that what we reconstruct as "schwa" represents different reduced vowels from the post-PIE period, incorrectly projected onto the PIE time plane.
 

3.  I know that there was no fixed form of a PIE infinitive, but what are 
some of the ways it could be formed and what would some of the PIE endings 
be?  

Most of the forms you find in various languages are datives (more rarely "directive" accusatives) of abstract nouns derived from verbal roots, with the meaning "for (the purpose of) V-ing" or "towards V-ing"; even the English infinitive is still used with to, which originated as a preposition accompanying the dative case. Slavic, for instance, has the ending *-ti < PIE *-tei(-ei) (the dative of action nouns in *-ti-s); a corresponding form also occurs in Baltic. Vedic retains several infinitives with the ending -ase < *-es-ei (the dative of neuter nouns in -os); a similar form (with certain complications) accounts for Latin -re (r is the regular Latin development of intervocalic *-s-, while the old consonant survives in esse 'to be'). In Classical Sanskrit the general infinitive ending was -tum (the accusative of nouns in *-tu-), formally corresponding to the Latin supine (expressing purpose with verbs of motion). Greek and Vedic have a variety of forms functioning as infinitives; the Celtic languages use various deverbal nouns, analysable as datives with the preposition do 'to'. In Germanic, the reconstructible ending of the infinitive is *-anan < PIE *-onom, probably another fossilised accusative of action nouns. For example, love is from OE lufian < PGmc *lubo:janan.
 

4.  What is the difference between athematic and thematic verbs?  (I don't 
even know if this is answerable because I have not seen a general consensus). 
 I have read that accent in the daughter languages (Sanskrit and Greek) is a 
sign of whether or not a PIE verb is thematic or athematic.  Is this true?  
And are there any other ways to tell?

WARNING: the distinction is valid only for non-Anatolian IE.
Thematic verbs have stems ending in a vowel; the vowel serves as a sort of buffer between the root and the inflectional endings, thus protecting the root from the phonetic influence of the ending. This helps to make the conjugation of thematic verbs highly regular and accounts for the great productivity of the thematic type. Athematic stems are consonant-final; their conjugation is often irregular. The productive type tends towards fixed stress, while the stress of the athematic type is usually mobile, falling on the root in the singular of the present tense, but shifting to the inflectional ending in the plural. Here are a couple of typical examples (in the present tense, with the stressed vowel capitalised):
bhEro-o, bhEre-si, bhEre-ti, bhEro-me-, bhEre-te-, bhEro-nti 'carry' (thematic)
 
ghwEn-mi, ghwEn-si, ghwEn-ti, ghw@..., ghw@..., ghwn-Onti 'strike' (athematic)
 
One almost foolproof criterion for telling thematic from athematic verbs is the form of the first person singular. The ending of thematic verbs is -o: versus the athematic -mi.
 

5.  Is the statement that "primary endings are used in the present indicative 
and secondary endings are used in the imperfect and aorist and in the 
optative, with fluctuation in the subjunctive" true?

Yes. It's not quite clear if there was any distinction in the 1st and 2nd person pl., but in the remaining sg. and pl. forms the primary endings had a final *i, probably expressing the present.
 

6.  In relation to primary and secondary endings, does a vowel always have to 
precede the ending?  If so, what determines the vowel (for example is it the 
outcome in specific daughter languages)?

 
A vowel appears in the thematic class; it's quality is *e before dental obstruents, and *o before a nasal and (whatever the reason) in the first person singular, e.g. bhereti vs. bherome- and bhero:. This looks like phonetic conditioning, though the details are not clear. The first person is often suspected of continuing a laryngeal ending, as in the Hittite -hi class of verbs or the 1st person sg. of the perfect. If that is correct, then dialectal *bhero: < *bhero-xa (with vowel contraction).
 
Very good questions! My best regards,
Piotr Gasiorowski